Research

The research focus of model development encompasses a multifaceted range of methods. Simple material and energy balances enable practical process and efficiency evaluation of anaerobic processes, while dynamic reaction models, considering microbial growth and characteristic intermediates or inhibitors, are suitable for model-based monitoring and control procedures. In addition to mechanistic modeling techniques, stochastic approaches based on machine learning are implemented within the working group as well. Long-standing experience in various modeling techniques allows for development of tailored model structures, depending on the individual objective and specific process conditions.

In addition to model development, analysis and validation of different process models also play a major role within the research group. Systematic investigations into parameter sensitivity, identifiability and observability enable a detailed and conclusive assessment of available model structures. Methods for optimal experimental design and parameter estimation at laboratory and industrial scales are implemented, evaluated and revised to enable clear identification of unknown model parameters and valuable process indicators.

The characterization of the utilized substrates is of decisive importance for reliable monitoring and precise simulation of anaerobic processes. Within the scope of the research group, available methods for substrate characterization are comparatively evaluated and specifically revised with regard to their application for process and efficiency evaluation. Model-based soft sensors for monitoring important state variables (process indicators) are developed for safe and flexible plant operation. In addition, many years of experience in manual process control can be utilized by means of expert systems for automated process monitoring. All methods are designed and evaluated for practical application with typical measurement uncertainties in large-scale plant operation.

Depending on the specific objective, model-based control procedures for different operating concepts of industrial anaerobic digestion plants are developed within the working group. Interdisciplinary collaboration, ranging from model development to condition monitoring and control theory, enables to design individual solutions for each specific application. In addition to the selection of suitable process models and numerical optimization procedures, the implementation of robust and predictive control methods is an essential prerequisite for successful application of model-based automation concepts in full-scale plant operation.

Experimental validation in laboratory and industrial scales is crucial for the development of practical automation concepts. Considering available sensors, laboratory analyses and typical measurement uncertainties, robust methods for process monitoring and control are developed and evaluated for regular plant operation. In this context, detailed assessment and continuous improvement of established or innovative measurement techniques and laboratory methods form an important focus within the research group. The goal is to develop or revise measurement methods that promise significant improvement and high informative value for model-based process optimization. Collaborations with plant manufacturers and operators provide valuable insights into workflows and operating conditions in practice.